Merge pull request #289 from ndaidong/7.0.2

v7.0.2

README.md

@@ -8,11 +8,12 @@ Extract main article, main image and meta data from URL.
 ![CodeQL](https://github.com/ndaidong/article-parser/workflows/CodeQL/badge.svg)
 [![JavaScript Style Guide](https://img.shields.io/badge/code_style-standard-brightgreen.svg)](https://standardjs.com)
 
+[![Deploy](https://button.deta.dev/1/svg)](https://go.deta.dev/deploy?repo=https://github.com/ndaidong/article-parser-deta)
 
 ## Demo
 
 - [Give it a try!](https://demos.pwshub.com/article-parser)
-- [Example FaaS](https://extractor.pwshub.com/article/parse?url=https://www.cnbc.com/2022/07/02/tesla-tsla-q2-2022-vehicle-delivery-and-production-numbers.html&apikey=demo-TEyRycuuMCiGBiBocbLGSpagfj7gOF8AMyAWfEgP)
+- [Example FaaS](https://extract-article.deta.dev/?url=https://www.freecodecamp.org/news/what-is-an-ide-for-beginners)
 
 ## Install
 

build.js

@@ -3,7 +3,7 @@
  * @ndaidong
  **/
 
-import { readFileSync, writeFileSync, rmSync, mkdirSync } from 'fs'
+import { readFileSync, writeFileSync, copyFileSync, rmSync, mkdirSync } from 'fs'
 
 import { buildSync } from 'esbuild'
 
@@ -75,3 +75,6 @@ const browserVersion = {
   }
 }
 buildSync(browserVersion)
+
+// copy types definition to cjs dir
+copyFileSync('./index.d.ts', 'dist/cjs/index.d.ts')

build.test.js

@@ -11,6 +11,7 @@ const pkg = JSON.parse(readFileSync('./package.json'))
 
 const cjsFile = `./dist/cjs/${pkg.name}.js`
 const cjsPkg = JSON.parse(readFileSync('./dist/cjs/package.json'))
+const cjsTypesFile = './dist/cjs/index.d.ts'
 
 describe('Validate commonjs version output', () => {
   test(`Check if ${cjsFile} created`, () => {
@@ -27,4 +28,7 @@ describe('Validate commonjs version output', () => {
     expect(lines[0].includes(pkg.author)).toBeTruthy()
     expect(lines[0].includes(pkg.license)).toBeTruthy()
   })
+  test(`Check if ${cjsTypesFile} created`, () => {
+    expect(existsSync(cjsTypesFile)).toBeTruthy()
+  })
 })

dist/cjs/index.d.ts

@@ -0,0 +1,92 @@
+// Type definitions
+
+import {AxiosRequestConfig} from "axios";
+import {IOptions as SanitizeOptions} from "sanitize-html";
+import {defaults} from "html-crush";
+import "urlpattern-polyfill";
+
+type HtmlCrushOptions = Partial<typeof defaults>
+
+/**
+ * @example
+ * {
+ *   patterns: [
+ *     '*://example.com/books/:id', {
+ *       hostname: 'example.com',
+ *       pathname: '/books/:id',
+ *     }
+ *   ],
+ *   selector: '.article-body',
+ *   unwanted: ['.removing-box']
+ * }
+ */
+export interface Transformation {
+  patterns: Array<URLPatternInit | string>,
+  pre?: (document: Document) => Document
+  post?: (document: Document) => Document
+}
+
+/**
+ * @param input url or html
+ */
+export function extract(input: string): Promise<ArticleData>;
+
+export function addTransformations(transformations: Array<Transformation>): Number;
+
+export function removeTransformations(options: Array<URLPatternInit>): Number;
+
+export function setParserOptions(options: ParserOptions): void;
+
+export function setRequestOptions(options: AxiosRequestConfig): void;
+
+export function setSanitizeHtmlOptions(options: SanitizeOptions): void;
+
+export function setHtmlCrushOptions(options: HtmlCrushOptions): void;
+
+export function getParserOptions(): ParserOptions;
+
+export function getRequestOptions(): AxiosRequestConfig;
+
+export function getSanitizeHtmlOptions(): SanitizeOptions;
+
+export function getHtmlCrushOptions(): HtmlCrushOptions;
+
+export interface ParserOptions {
+  /**
+   * For estimating "time to read".
+   * Default: 300
+   */
+  wordsPerMinute: number
+  /**
+   * To find the best url from list
+   */
+  urlsCompareAlgorithm: 'levenshtein' | 'cosine' | 'diceCoefficient' | 'jaccardIndex' | 'lcs' | 'mlcs'
+  /**
+   * Min num of chars required for description
+   * Default: 40
+   */
+  descriptionLengthThreshold: number
+  /**
+   * Max num of chars generated for description
+   * Default: 156
+   */
+  descriptionTruncateLen: number
+  /**
+   * Min num of chars required for content
+   * Default: 200
+   */
+  contentLengthThreshold: number
+}
+
+export interface ArticleData {
+  url?: string;
+  links?: string[];
+  title?: string;
+  description?: string;
+  image?: string;
+  author?: string;
+  content?: string;
+  source?: string;
+  published?: string;
+  ttr?: number;
+}

dist/cjs/package.json

@@ -1,5 +1,5 @@
 {
   "name": "article-parser-cjs",
-  "version": "7.0.1",
+  "version": "7.0.2",
   "main": "./article-parser.js"
 }

package.json

@@ -1,5 +1,5 @@
 {
-  "version": "7.0.1",
+  "version": "7.0.2",
   "name": "article-parser",
   "description": "To extract main article from given URL",
   "homepage": "https://ndaidong.github.io/article-parser-demo/",
@@ -30,12 +30,12 @@
   "dependencies": {
     "@mozilla/readability": "^0.4.2",
     "axios": "^0.27.2",
-    "bellajs": "^11.0.4",
-    "html-crush": "^5.0.22",
+    "bellajs": "^11.0.5",
+    "html-crush": "^5.1.3",
     "linkedom": "^0.14.12",
     "sanitize-html": "^2.7.1",
     "string-comparison": "^1.1.0",
-    "tldts": "^5.7.89",
+    "tldts": "^5.7.90",
     "urlpattern-polyfill": "^5.0.6"
   },
   "standard": {
@@ -46,8 +46,8 @@
   "devDependencies": {
     "@types/sanitize-html": "^2.6.2",
     "cross-env": "^7.0.3",
-    "esbuild": "^0.15.1",
-    "jest": "^28.1.3",
+    "esbuild": "^0.15.6",
+    "jest": "^29.0.1",
     "nock": "^13.2.9",
     "standard": "^17.0.0"
   },